In The Hot Seat: ATI's Phil Rogers

[Ubermann]

In 1999, Rogers moved to an architectural role and has been responsible for 3D driver architecture and 3D performance optimization for the Radeon series of GPUs. An important part of that role has been to work closely with Microsoft on the evolution of the Direct3D API/DDI and on the development of the Windows Display Driver Model for Windows Vista. Rogers plays an active role helping Microsoft make software architecture decisions that work well with all AMD graphics hardware and drivers.

In 2006 Rogers was promoted to ATI Fellow with broad architectural responsibility for all graphics software at ATI, and now AMD. In this role Rogers has guided development of AMD's Vista drivers, HD-DVD, and Blu-Ray playback drivers, a new driver architecture for Dx10, and the Crossfire architecture for MultiGPU. In his spare time Rogers directed development of the new 3D preview for Catalyst Control Center, which just released on Vista.


Check the Interview at ExtremeTech (http://www.extremetech.com/article2/0,1697,2103630,00.asp)

[perkam]

Nicee.

Perkam

[Cybercat]

Ah ha! So Microsoft is biased towards AMD with their OS. No wonder AMD has an easier time releasing stable drivers.

[deathman20]

Ah ha! So Microsoft is biased towards AMD with their OS. No wonder AMD has an easier time releasing stable drivers.

Well AMD does have a GPU in there Console. So they will be a little more biased to them of course because im sure they've been working closer with AMD then nVidia.

[citat3962]

If this means stable drivers and better integration with Vista I think ATI may have an ace in the hole even after the ridiculously LONG wait.

[nn_step]

Ah ha! So Microsoft is biased towards AMD with their OS. No wonder AMD has an easier time releasing stable drivers.

well Every OS is in some way Biased for certain hardware. The same goes for every single application. For example lets just say your entire application runs in 16 registers, thus in a Processor with 16 Registers you get optimal performance, while in a 8 register Processor you can get some serious penalties by swapping the register contents in and out.

[Cybercat]

Be that as it may, it's still all deliciously scandalous. :devil: This was my favorite part:

Our Windows XP graphics memory manager is called VCAM. In Windows Vista, the graphics memory manager is a Microsoft component called VidMM, used by all IHVs. So we worked closely with Microsoft on the design of VidMM, to ensure it offered all the features of our own VCAM and helped Microsoft debug and tune it. As a result, we find VidMM to be very good for AMD hardware and it allows our drivers the same flexibility in placing our surfaces for optimal performance. Indeed we are able to use common code in our unified driver for operating VidMM in our Vista drivers and VCAM in our Windows XP drivers.

Bam, right there, they automatically get the edge in stability/compatibility. NVIDIA now has to adapt their memory manager to be more like ATI's, which means problems for them.

[eXceededgoku]

^^ about time ATI wasnt the underdog... Nvidia needs to catch up on software, whereas ATI needs to catch up on hardware!

[Iconyu]

ATi have been close to Microsoft for some time, infact since before the 9700's. They had a lot of input with DX9.0, and (in my opinion) it was the deal with Microsoft for the 360 that stopped them from releasing a SM3.0 chip when nvidia's 6800's came out. The silicon they had ready was held in standby meaning they had to convert R300 to do the job.

[Shadowmage]

well Every OS is in some way Biased for certain hardware. The same goes for every single application. For example lets just say your entire application runs in 16 registers, thus in a Processor with 16 Registers you get optimal performance, while in a 8 register Processor you can get some serious penalties by swapping the register contents in and out.

What are you smoking? x86 and its other extensions dictate a set amount of architectural registers.

[nn_step]

What are you smoking? x86 and its other extensions dictate a set amount of architectural registers.

Comparing ATi to nVidia is like comparing PowerPC and x86 (yes they are that architecturally different) Not like comparing AMD vs intel x86(_64) Processors

[Shadowmage]

Comparing ATi to nVidia is like comparing PowerPC and x86 (yes they are that architecturally different) Not like comparing AMD vs intel x86(_64) Processors

Sorry, wrong again.

http://www.behardware.com/articles/631-2/directx-10-and-gpus.html

Like the ISA for a CPU, DirectX 10 specifies the number of architected registers. Note that ATI and NVIDIA both have the same number of registers in the table.

[ColonelCain]

This is old news to me. Except, now I know the mastermind behind the ideas.

[P_1]

The guy doing the interview is an idiot, he kept asking the same question trying to make Phil Rogers say that Vista is slower in games when it is not. The interviewer clearly does not know what is going on at all, because all of the latest benchmarks for AMD/ATI drivers show that it is faster then xp in almost every case. Yet this guy keeps pushing for the response that he wishes to hear(that vista is slower than xp, when it is not).

[nn_step]

Sorry, wrong again.

http://www.behardware.com/articles/631-2/directx-10-and-gpus.html

Like the ISA for a CPU, DirectX 10 specifies the number of architected registers. Note that ATI and NVIDIA both have the same number of registers in the table.

I'll give you a shiny penny if you can right one program using ATi's CTM software specification and get it to run on ANY nVidia GPU. Because you can right any application in x86 assembly and it'll work on ANY AMD or Intel x86 or x86_64 Processor (provided you follow the x86 Specification)

[Shadowmage]

I'll give you a shiny penny if you can right one program using ATi's CTM software specification and get it to run on ANY nVidia GPU. Because you can right any application in x86 assembly and it'll work on ANY AMD or Intel x86 or x86_64 Processor (provided you follow the x86 Specification)

That doesn't matter. My whole point was that your register example was completely and utterly incorrect.

[Quest_7F]

^^^^^^Why would you point that out, why not add or correct NN. Instead you ruin the very aspect of a discussion.